On Thu, 02 Apr 2026 15:48:01 +0200,
Visa Hankala <visa@hankala.org> wrote:
> 
> On Wed, Apr 01, 2026 at 09:44:40PM +0200, Kirill A. Korinsky wrote:
> > Visa, tech@,
> > 
> > in my work to bring support queues in cnmac I think I discovered a weird
> > issue in interrupts and barrier in octeon.
> > 
> > octciu_intr_establish() has:
> > 
> >    201          int cpuid = cpu_number();
> > ...
> >    208  #ifdef MULTIPROCESSOR
> >    209          /* Span work queue interrupts across CPUs. */
> >    210          if (IS_WORKQ_IRQ(irq))
> >    211                  cpuid = irq % ncpus;
> >    212  #endif
> > ...
> >    222          ih->ih_irq = irq;
> > 
> > my understanding is that octciu_intr_establish() happens on autoconf where
> > only CPU0 exists, and on octeon we increase ncpus in cpu hatch, and not
> > attach, which means that ncpus is 1 at octciu_intr_establish() time.
> > 
> > What makes cpuid always 1 for that code and schedules all interrupts on
> > CPU0.
> 
> Hmm, the code used ncpusfound originally. But it was later changed to
> use ncpus, which apparently broke the logic.
>

Thanks for the lead. This reads like regression which I see with kn@'s diff
https://github.com/openbsd/src/commit/fee69528d2cd998356813db0922e7fd576485f9c

> > Next, octciu_intr_barrier():
> > 
> >    290  void
> >    291  octciu_intr_barrier(void *_ih)
> >    292  {
> >    293          struct cpu_info *ci = NULL;
> >    294  #ifdef MULTIPROCESSOR
> >    295          struct octciu_intrhand *ih = _ih;
> >       
> >    296          if (IS_WORKQ_IRQ(ih->ih_irq))
> >    297                  ci = get_cpu_info(ih->ih_irq % ncpus);
> >    298  #endif
> >       
> >    299          sched_barrier(ci);
> >    300  }
> > 
> > because ih->ih_irq is 0 here, it picks cpu0 only.
> > 
> > kn@'s diff https://marc.info/?l=openbsd-bugs&m=177171071006897&w=2
> > introduced a regression on my ER-4:
> > https://marc.info/?l=openbsd-bugs&m=177177669428825&w=2 and this regression
> > can be avoided by removing the block in octciu_intr_establish().
> > 
> > Here is the diff where I remove this block and % ncpus.
> > 
> > This diff keeps interrupts on CPU0 which is, actually, current behaviour and
> > unblocks kn@'s diff to enable softnet threads on octeon which unblocks my
> > work of making cnmac with RX queue:
> > https://marc.info/?l=openbsd-tech&m=177506394121597&w=2
> > 
> > So far I see this diff as safe because it doesn't change current behaviour
> > but I may miss some details.
> > 
> > OK to commit it?
> 
> I suggest you surround the #ifdef MULTIPROCESSOR blocks in
> octciu_intr_establish() and octciu_intr_barrier() with #if 0, because
> you will probably need them later for distributing the POW group
> interrupts and testing your multi-queue work.
> 

I still not completley sure that the second block in octciu_intr_barrier()
need to be removed.

In my full diff, which I send into tech@ as cumulative one I don't touch
octciu_intr_barrier(), and I run it on my home's gw and it works well,
stable, and I have near ~1gbps between vlans on the same cnmac.

So, here a diff which I think is right, which shouldn't change current
behaviour, and which avoid regression from kn@'s difs and on my gw with two
diffs I see multiple softnet threads and don't see the regression.

Ok?

Index: sys/arch/octeon/dev/octciu.c
===================================================================
RCS file: /home/cvs/src/sys/arch/octeon/dev/octciu.c,v
diff -u -p -r1.19 octciu.c
--- sys/arch/octeon/dev/octciu.c	11 Dec 2022 05:31:05 -0000	1.19
+++ sys/arch/octeon/dev/octciu.c	2 Apr 2026 21:27:14 -0000
@@ -250,11 +250,13 @@ octciu_intr_establish(int irq, int level
 		panic("%s: illegal irq %d", __func__, irq);
 #endif
 
+#if 0
 #ifdef MULTIPROCESSOR
 	/* Span work queue interrupts across CPUs. */
 	if (IS_WORKQ_IRQ(irq))
 		cpuid = irq % ncpus;
 #endif
+#endif
 
 	flags = (level & IPL_MPSAFE) ? CIH_MPSAFE : 0;
 	level &= ~IPL_MPSAFE;
@@ -363,7 +365,7 @@ octciu_intr_barrier(void *_ih)
 	struct octciu_intrhand *ih = _ih;
 
 	if (IS_WORKQ_IRQ(ih->ih_irq))
-		ci = get_cpu_info(ih->ih_irq % ncpus);
+		ci = get_cpu_info(ih->ih_cpuid);
 #endif
 
 	sched_barrier(ci);


-- 
wbr, Kirill